Perturbation Effects on Robustness and Individual Fairness

Xuran Li, Hao Xue, Peng Wu +4

Deep neural networks are vulnerable to adversarial perturbations that can simultaneously degrade prediction robustness and individual fairness across diverse application settings.…